Exploring the Relationship between Paleobiogeography, Deep-Diving Behavior, and Size Variation of the Parietal Eye in Mosasaurs. Andrew M.

Transcription

1 Exploring the Relationship between Paleobiogeography, Deep-Diving Behavior, and Size Variation of the Parietal Eye in Mosasaurs By Andrew M. Connolly Submitted to the graduate degree program in Geology and the Graduate Faculty of the University of Kansas in partial fulfillment of the requirements for the degree of Master of Arts. Stephen T. Hasiotis, Chairperson Rafe M. Brown Jennifer A. Roberts Date Defended: March 25, 2016 i

2 The Thesis Committee for Andrew M. Connolly certifies that this is the approved version of the following thesis: Exploring the Relationship between Paleobiogeography, Deep-Diving Behavior, and Size Variation of the Parietal Eye in Mosasaurs Stephen T. Hasiotis, Chairperson Date Approved: March 25, 2016 ii

3 ABSTRACT Andrew M. Connolly, M.S. Department of Geology, March 2015 University of Kansas The parietal eye (PE) in modern squamates (Reptilia) plays a major role in regulating body temperature, maintaining circadian rhythms, and orientation via the solar axis. This study is the first to determine the role, if any, of the PE in an extinct group of lizards. We analyzed variation in relative size of the parietal foramen (PF) of five mosasaur genera to explore the relationship between PF size and paleolatitudinal distribution. We also surveyed the same specimens for the presence of avascular necrosis a result of deepdiving behavior in the vertebrae. Plioplatecarpus had the largest PF followed by Platecarpus, Tylosaurus, Mosasaurus, and Clidastes. A weak relationship exists between paleolatitudinal distribution and PF size among genera, as Plioplatecarpus had the highest paleolatitudinal distribution (~78 N) and the largest PF among genera. Clidastes, Mosasaurus, Platecarpus, and Tylosaurus, however, shared a similar northern paleolatitude (~55 N) extent despite Platecarpus having a statistically larger PF than the other three genera (p<0.001 in Fisher s LSD test). Mosasaurus, Plioplatecarpus, and Tylosaurus also shared a similar southern paleolatitude (~64 S) despite Plioplatecarpus having a larger PF. There is no correlation between PF size and paleolatitudinal distribution for specimens within genera. We found no relationship between PF size and presence of avascular necrosis. Tylosaurus and Mosasaurus, which exhibited avascular necrosis, had a similar PF size to Clidastes, which did not avascular necrosis. The PE of mosasaurs may have functioned primarily for navigation and orientation related to migration; however, this possibility requires further study of modern PE-bearing organisms and its function. iii

4 ACKNOWLEDGEMENTS First, I would like to thank everyone who came and listened to me talk about my research for these past few years. A variety of strangers have given me help from simple words of encouragement to very sound advice. Likewise, an extra thanks goes to all the mosasaur experts out there, especially to those who saw me present at the International Mosasaur Conference. Thank you to Dr. Konishi for reidentifying a key specimen that was found in Northwest Territories fifty years ago and was almost completely forgotten since. Thank you to Dr. Everhart and Dr. Polycn for answering the numerous questions I had about mosasaurs. Thank you to Gil Parker, president of the Kansas City Paleontology Club, for calling the Black Hills Institute in South Dakota and speaking to them on my behalf for a visit that was planned on short notice. Thank you to various museums and collection curators who allowed me to visit their collections and photograph their mosasaur specimens. Thank you to my graduate colleagues for these past few years, especially to IchnoBioGeoScience, who taught me how to write well (e.g., the oxford comma, when to abbreviate a species name, and how to use an e.g. ) and write decent scientific papers. Thanks especially to Dr. Amanda Falk, Dr. Joshua Schmerge, Matthew Jones, Tabitha Gabay, Nicole Dznowski, Matthew Downen, and Derek Raisanan. The list of things you guys did for me would be too long to write here so I ll simply say that you guys are awesome. Although not directly affecting my thesis, my lab supervisor, Julie Campbell, should also be thanked as she continuously hired me to be a Biology Lab TA. Her positive outlook, humorous attitude, and love of education are quite infectious. You are a definite morale iv

5 booster and I don t have any student loans because of you! Also, I want to thank my students I had for the past several semesters. You guys taught me that I love teaching. When times were troubling, I could always look forward to you guys and teach another fun and wonderful lesson knowing that you care about science as much as I do. You guys were worth it. Thank you to my Dad for driving me to various museums on our family vacation. I know that the Field Museum and especially the Yale Peabody Museum were out of the way, but you sacrificed the time and the money to ensure I could visit these outstanding museums. All because you wanted me to succeed and become a great scientist. Thanks Dad, I really, really, appreciate that. Thank you to the rest of my family as well for, again, being my moral support. Thank you to my committee members, especially Dr. Rafe Brown who filled in as my third committee member after Dr. Larry Martin past away, and of course thank you to Dr. Stephen Hasiotis. You took me on as my advisor when Larry Martin past away and gave me moral support and scientific advice. That was a big task that I could hardly imagine few others would be willing to take. Although I learned a lot from my graduate courses, you and your students have taught me so much in how to write, research, and be a respectable scientist. This was worth more to my experience as a scientist than all my courses combined. And almost lastly, Dr. Larry Martin. Who got me started on this project and had faith in me in not only as a scientist and a researcher but even as a good public speaker. His encouragement for me to apply to my first conference, the Kansas Academy of Sciences, in 2011 awarded me 2 nd place in best undergraduate oral presentation. From there, my v

6 confidence in myself and what I was doing soared and encouraged me to apply to other conferences even after his passing. We miss you and I still wish you were here to give me your unmatched paleontological advice. And lastly, Mary Fluker, my girlfriend who was with me since just before I started graduate school at the University of Kansas. You accompanied me on my research trips and gave me company in what otherwise would be long and lonely drives. Thank you so much for you support and I love you. vi

7 TABLE OF CONTENTS ABSTRACT...iii ACKNOWLEDGEMENTS...iv FIGURES AND TABLES...viii CHAPTER ONE. EXPLORING THE RELATIONSHIP BETWEEN PALEOBIOGEOGRAPHY, DEEP-DIVING BEHAVIOR, AND SIZE VARIATION OF THE PARIETAL EYE IN MOSASAURS 1 REFERENCES.28 APPENDIX..34 vii

8 FIGURES AND TABLES Figures Figure 1. Dorsal view of a Plioplatecarpus skull 3 Figure 2. Dorsal view of mosasaur parietal bones..6 Figure 3. A map of mosasaur fossil localities that were analyzed for the experiment 9 Figure 4. Graph showing the average parietal foramen size among genera..14 Figure 5. Graphs showing the relationship between relative parietal foramen size and paleolatitudinal distribution...16 Figure 6. A generalized mosasaur phylogenetic tree.17 Figure 7. Plot of the residual independent contrasts between parietal foramen length and parietal bone length...17 Tables Table 1. Descriptions for Hypothesis One and Hypothesis Two 7 Table 2. Measurements and locality information for all mosasaur specimens by species..9 Table 3. Highest known northern and southern paleolatitudinal distribution for each mosasaur genera 14 Table 4. : Presence or absence of avascular necrosis in each mosasaur genera 18 Appendices Appendix 1: Relative parietal foramen size among different lizards as arranged by species...35 viii

9 1 2 3 CHAPTER 1. EXPLORING THE RELATIONSHIP BETWEEN PALEOBIOGEOGRAPHY, DEEP-DIVING BEHAVIOR, AND SIZE VARIATION OF THE PARIETAL EYE IN MOSASAURS Currently in review as: Connolly, A.M., Hasiotis, S.T., and Martin, L.D., Exploring the relationship between paleobiogeography, deep-diving behavior, and size variation of the parietal eye in mosasaurs. Journal of Vertebrate Paleontology Parietal Foramen, Plioplatecarpus, Platecarpus, Avascular Necrosis, Mosasaurus ABSTRACT The parietal eye (PE) in modern squamates (Reptilia) plays a major role in regulating body temperature, maintaining circadian rhythms, and orientation via the solar axis. This study is the first to determine the role, if any, of the PE in an extinct group of lizards. We analyzed variation in relative size of the parietal foramen (PF) of five mosasaur genera to explore the relationship between PF size and paleolatitudinal distribution. We also surveyed the same specimens for the presence of avascular necrosis a result of deepdiving behavior in the vertebrae. Plioplatecarpus had the largest PF followed by Platecarpus, Tylosaurus, Mosasaurus, and Clidastes. A weak relationship exists between paleolatitudinal distribution and PF size among genera, as Plioplatecarpus had the highest paleolatitudinal distribution (~78 N) and the largest PF among genera. Clidastes, Mosasaurus, Platecarpus, and Tylosaurus, however, shared a similar northern paleolatitude 1

10 (~55 N) extent despite Platecarpus having a statistically larger PF than the other three genera (p<0.001 in Fisher s LSD test). Mosasaurus, Plioplatecarpus, and Tylosaurus also shared a similar southern paleolatitude (~64 S) despite Plioplatecarpus having a larger PF. There is no correlation between PF size and paleolatitudinal distribution for specimens within genera. We found no relationship between PF size and presence of avascular necrosis. Tylosaurus and Mosasaurus, which exhibited avascular necrosis, had a similar PF size to Clidastes, which did not avascular necrosis. The PE of mosasaurs may have functioned primarily for navigation and orientation related to migration; however, this possibility requires further study of modern PE-bearing organisms and its function INTRODUCTION The importance of the parietal eye (PE), also known as the pineal eye, in mosasaurs has yet to be determined as no studies have examined its function in this ancient group of marine reptiles. Additionally, no studies have examined the function of the PE in any extinct group of animals so this study is the first to determine if ancient animals used their PE similarly to their modern counterparts. Today, the PE is found only in some frogs, lampreys, Lepidosauria (squamates plus Sphenodon), salamanders, and toads (Tosini, 1997). The PE is located on the dorsal, posterior part of the skull and can be between the two frontal bones, the two parietal bones (PB), or in the frontoparietal suture (Fig. 1). The PE possesses a cornea, retina, and lens; however, the PE can only sense the presence of light (Edinger, 1955). The parietal foramen (PF) is the orifice in the skull that holds the PE, which is connected to the pineal gland. Collectively, the PE, PF, and pineal body comprise the pineal complex. 2

11 Figure 1. Dorsal view image of a Plioplatecarpus skull. Frontal bone (f), parietal foramen (pf) and parietal bone (p) are labeled. Image courtesy of Oceans of Kansas website at The primary function of the pineal complex is maintaining circadian rhythms, body temperature, and orientation via the solar axis (Ralph, 1975). In particular, the pineal complex generally becomes more pronounced and larger for animals living in high (and generally cooler) latitudes, possibly as a result of selection for increased sensitivity to lower intensity sunlight (Ralph, 1975; Quay 1980). For example, Davenport et al. (2014) proposed that the pineal body triggers migration in the leatherback sea turtle, Dermochelys coriacea, due to seasonal variation in day length. This turtle has a thin PB referred to as a skylight by Davenport et al. (2014) which allows greater absorption of the low intensity sunlight at higher latitudes by the pineal body. Dermochelys also migrates seasonally to high latitudes, such as Newfoundland (~49 N) (Goff et al., 1988), Scotland (~55 N) 3

12 (Davenport et al., 2014), and northern Norway (~71 N) (Carriol and Vader, 2002) before they return to their Caribbean nesting grounds. Davenport et al. (2014) noted that no other group of migratory sea turtles possesses a similarly modified PB, nor do they migrate to as high of latitude as D. coriacea. Gundy et al. (1975), likewise, found that most agamid (Squamata: Agamidae) and iguanid (Squamata: Iguanidae) lizards lacking a PE are restricted to 10 of the equator, whereas PE-bearing lizards of the same genera are found up to 55 from the equator. Gundy et al. (1975) concluded that PE-bearing lizards are better adapted to the more extreme daily and seasonal temperature variations of higher latitudes than those that do not have a PE. A biogeographical study of lizard skulls in the University of Kansas Herpetology Collection found similar results, as such temperate-latitude lizards as Holbrookia maculata (Squamata: Phrynosomatidae) have a larger PF a proxy for the PE than such equatorial lizards as Stenocercus variabilis (Squamata: Tropiduridae) (Appendix 1). Labra et al. (2010), however, found no relationship between latitudinal or altitudinal distribution or local environmental temperatures to PE size in 30 species of the South American lizard Liolaemus (Squamata: Liolaemidae). These contradictory results complicate our current understanding of the function of the PE in lizards in terms of thermoregulation and circadian rhythms. Lizards also use their PE to orient themselves relative to their environment (e.g., Gundy et al., 1975, Freake, 2001, Beltrami et al., 2010). Beltrami et al. (2010) demonstrated that the ruin lizard, Podarcis sicula (Squamata: Lacertidae), orients spatially via an internal compass and reference to solar cues. When the PE of test subjects was experimentally ablated, orientation performance declined. Freake (2001) displaced individuals of the Australian sleepy lizard, Tiliqua rugosa (Squamata: Scincidae), 800 m 4

13 from their home range, and found that only individuals with a functional PE returned to their home range. This nonmigratory species would also occasionally leave their home range for reproduction, employing olfaction and visual cues (presumably including solar cue detection via the PE). Similarly, Ellis-Quinn and Simon (2004) found that specimens of Yarrow s spiny lizard, Sceloporus jarrovi, had a significantly higher percentage of individuals returning to their natural home range (after being displaced 150 m away) with an uncovered PE compared to those with a covered PE. Although many experiments have been conducted to understand the function of the PE in extant lizards, no studies have explored the function of the PE using the PF as its proxy in ancient squamates. We studied the relative size of the PE in five mosasaur genera to determine whether the PE in an extinct group of lizards functioned similarly to that of their extant relatives. Mosasaurs (Squamata: Mosasauridae), a group of marine lizards that lived during the Late Cretaceous Period, may have used their PE to help regulate their circadian rhythms, maintain body temperature, or orient themselves relative to sea-penetrated sunlight. Mosasaurs are an excellent group of ancient organisms to determine the function of their PE. There is an obvious difference in the relative size of their PF (Fig. 2) and they have a global distribution (e.g., Russell, 1967, Kear et al., 2005, Páramo-Fonseca, 2011, Fernández and Gasparini, 2012, Leblanc et al., 2012), both of which could help resolve if PE size and paleolatitudinal distribution correlate. What is more difficult to determine is if mosasaurs used their PE for orientation as we cannot study their behavior directly relative to the rays of the sun. We can analyze, however, their bone structures to determine what diving behavior each mosasaur genera had. Some mosasaur genera (e.g., Platecarpus, Plioplatecarpus, Tylosaurus) were probably deep-diving predators based on the presence of 5

14 avascular necrosis bone tissue that has very porous spongy structure produced by degassing within the bone while ascending from deep to shallow water (i.e., the bends) in the vertebrae (Rothschild and Martin, 2005). Perhaps these deep-diving mosasaurs had a large, light-sensitive PE which could orient them in their low-light environment. Shallowdwelling mosasaurs, likewise, would not need a pronounced PE due to the high amount of light stimuli in their environment

15 Figure 2. Dorsal view of mosasaur parietal bones in A, Mosasaurus (RBINS VF AR 42-12); B, Platecarpus (SMU 76351); C, Plioplatecarpus (RMM 7071); and D, Tylosaurus (RMM 5610). Scale bar equals 5 cm. As such, we propose two hypotheses (H1, H2; Table 1) to explain the function of the PE in mosasaurs: H1) PF size increases as paleolatitudinal distribution of fossil specimens increases; and H2) PF size is relative to the amount of avascular necrosis in mosasaur vertebrae. For H1 and H2, the PF length is used as a proxy for PE size. For H1, we compare the PF size to paleolatitudinal distribution among and within mosasaur genera. Hypothesis 1 is falsified if there is no relationship between PF size and paleolatitudinal distribution. For H2, a larger PF is associated with a greater number of vertebrae affected by avascular necrosis. Hypothesis 2 is falsified if there is no relationship between PF size and amount of avascular necrosis in the mosasaur vertebrae. 126 Hypotheses H1 H2 Description Size of parietal foramen increases for mosasaurs in higher paleolatitudes, both within and among genera Size of parietal foramen is larger in mosasaurs evidence of avascular necrosis compared to mosasaurs without it. 127 Table 1. Descriptions for Hypothesis One (H1) and Hypothesis Two (H2) Institutional Abbreviations ALMNH PV, the Alabama Museum of Natural History, Tuscaloosa, Alabama, U.S.A.; FHSU, the Fort Hays Sternberg Museum, Hays, Kansas, U.S.A.; FMNH, the Field Museum of Natural History, Chicago, Illinois, U.S.A.; ING, Servicio Geológico Colombiano, Bogota, Columbia; KHM, Kaikoura Historical Museum, Kaikoura, New Zealand; KUVP, the University of Kansas Museum of Natural History, Lawrence, Kansas, U.S.A.; M, Canadian Fossil Discovery Centre (previously 7

16 Morden and District Museum), Morden, Manitoba, Canada; MLP, Museo de La Plata, Buenos Aires, Argentina, MNHN, the Muséum National D'histoire Naturelle, Paris, France; NMC, Canadian Museum of Nature, Ottawa, Ontario; NMNZ, the National Museum of New Zealand, Wellington, New Zealand; OCP, Office Chérifien des Phosphates, Khouribga, Morocco; RBINS, the Institut Royal des Sciences Naturelles de Belgique, Brussels, Belgium; RMM, the Mcwane Center, Birmingham, Alabama, U.S.A.; SMU, the Shuler Museum of Paleontology at the Southern Methodist University, Dallas, Texas, U.S.A; TMP, the Royal Tyrell Museum of Paleontology, Drumheller, Alberta, Canada; UALVP, University of Alberta Laboratory for Vertebrate Paleontology, Calgary, Alberta, Canada; UNO, University of New Orleans, New Orleans, Louisiana, U.S.A.; USNM, National Museum of Natural History, Washington D.C., U.S.A.; WDC, Wyoming Dinosaur Center, Thermopolis, Wyoming, U.S.A.; YPM, Yale Peabody Museum, New Haven, Connecticut, U.S.A MATERIALS AND METHODS Material We examined five mosasaur genera: Clidastes (n=13); Mosasaurus (n=5); Platecarpus (n=13); Plioplatecarpus (n=10); and Tylosaurus (n=10). We chose these taxa because of their: (1) variation in PF size (see Fig. 2); (2) paleolatitudinal distribution (Fig. 3); and (3) abundance in various fossil collections in North America and in Europe (Table 2). We took special effort to measure nonkansan mosasaurs to ensure a diverse paleolatitudinal range among the five genera, as Kansan mosasaurs are found in museums worldwide. We only included data from mosasaurs with a well-preserved PB. We measured 8

17 other mosasaur genera, such as Prognathodon and Tethysaurus (see Table 2), but their sample size was too small to be included in these analyses Figure 3. General location of all mosasaur specimens measured in the northern hemisphere. 1) Mosasaurus; 2) Platecarpus; 3) Clidastes, Mosasaurus, Platecarpus; 4) Clidastes, Plioplatecarpus, Tylosaurus; 5) Platecarpus, Tylosaurus; 6) Clidastes, Plioplatecarpus, Tylosaurus; 7) Mosasaurus, Plioplatecarpus. 165 Catalogue # Species Name Location YPM 1333 ALMNH PV YPM 1368 Clidastes liodontus Clidastes liodontus "moorevillensis" Clidastes propython Length of Parietal Foramen (PF) (mm) Length of Parietal Bone (PB) (mm) PF/PB Kansas: Graham County, Solomon River Alabama: Greene County Kansas: Wallace County, 23 miles east of Fort Wallace

18 FSHU FMNH P FMNH PR 495 ALMNH PV RMM 2423 TMP Clidastes propython Clidastes propython Clidastes propython Clidastes propython Clidastes propython Clidastes propython ALMNH PV Clidastes sp. ALMNH PV Clidastes sp. Kansas: Logan County, Coal Oil Canyon Alabama: Dallas County, 3 miles southeast of Harrell Station Kansas: Logan County Alabama: Greene County Alabama: Greene County South Dakota: Pennington County Alabama: Greene County Alabama: Greene County USNM-PAL Clidastes sp. Kansas MNHN PMC 14 Halisaurus arambourgi Morroco: Near Grand Daoui area WDC CPM- 100 USNM SMU PA 18 FMNH PR 467 TMP KUVP 1034 RBINS VF AR 42/12 Halisaurus arambourqi Morroco Maryland: Prince Georges Halisaurus County, Oxon platyspondylus Hill Halisaurus sp. Latoplatecarpus nichollsae Latoplatecarpus willistoni Mosasaurus horridus Mosasaurus lemmonnieri Angola: Bentiaba South Dakota: Custer County Manitoba: Morden South Dakota: Custer County Belgium: Hainaut Province, Ciply

19 RBINS VF AR 42/15 RBINS VF AR 42/8 Mosasaurus lemmonnieri Mosasaurus lemmonnieri TMP Mosasaurus sp. YPM 1112 YPM 3690 M YPM USNM M M KUVP 1085 SMU FHSU RBINS 3100 RBINS 3101 RBINS 3108 RBINS R 36 Platecarpus ictericus Platecarpus ictericus Platecarpus sp. Platecarpus sp. Platecarpus sp. Platecarpus tympaniticus Platecarpus tympaniticus Platecarpus tympaniticus Platecarpus willistoni Plesioplatecarpus planifrons Plioplatecarpus houzeaui Plioplatecarpus houzeaui Plioplatecarpus houzeaui Plioplatecarpus houzeaui Belgium: Hainaut Province, Spiennes Belgium: Hainaut Province, Ciply, Alberta: South of Lethbridge Kansas: Logan County, Russell Springs Kansas: Gove or Lane County Manitoba: Pembina Escarpment Kansas: Wallace County Kansas: Logan County near Scott City Manitoba: Pembina Escarpment Manitoba: Pembina Escarpment Kansas: Logan County, Hell Creek North Central Texas Kansas: Gove County, Tuffle Ranch Belgium: Hainaut Province, Ciply Belgium: Hainaut Province, Ciply Belgium: Hainaut Province, Ciply Belgium: Hainaut Province, Ciply

20 RBINS R 40 NMC NMC KU UNO RMM 7071 SMU PA 149 RBINS R33 SMU FHSU MNHN GOU2 USNM PAL FHSU FHSU 2209 KB-MC-M- 16/SMU Triebold "Sophie" ALMNH PV RMM 5610 Plioplatecarpus houzeaui Plioplatecarpus primaevus Plioplatecarpus primaevus Plioplatecarpus sp. Plioplatecarpus sp. Prognathodon kianda Prognathodon solvayi Russellosaurus coheni Selmasaurus johnsoni Tethysaurus nopcsai Belgium: Limburg Province, Kanne South Central Saskatchewan South Central Saskatchewan Alabama: Coatopa County Kansas: Unknown, in Mooreville Chalk, possibly Greene County Angola: Bentiaba Belgium: Mons Basin Texas: Dallas County: Cedar Hills Kansas: Gove County Morocco: Goumima region Tylosaurus dyspelor Kansas Tylosaurus kansensis Tylosaurus nepaeloicus Tylosaurus nepaeolicus Tylosaurus proriger Tylosaurus proriger Tylosaurus proriger Kansas: Gove County Kansas: Rooks County Texas: Brewster County, 10 km southwest of Terlingua Ghost Town Texas: Ellis County Alabama: Greene County Alabama: Hale County

21 TMP Tylosaurus proriger Kansas: Graham County YPM 3990 Tylosaurus proriger Kansas: Wallace County YPM 3993 Tylosaurus proriger Kansas: Wallace County Table 2. Measurements recorded for the parietal foramen (PF) length, the parietal bone (PB) length, and the PF/PB for every specimen along with its excavated locality. Note that some locality data are incomplete, such as USNM PAL , and were not used for testing H2 within genera Methods Each skull was photographed using a Sony Cyber-shot DSC-H MP Digital Camera. We analyzed each photograph with J. Microvision v1.27 to obtain exact measurements of the PB and PF length (see Fig. 1), and calculated the relative size of the PF by comparing the length of the PF to the length of the PB (i.e., PF/PB). We needed to measure the relative size of the PF as large mosasaurs, like Tylosaurus (see Fig. 2), would naturally have a larger PF compared to smaller mosasaurs like Clidastes. As such, the PB was used as it gave us a high amount of measured individuals and ensured that PF size would not be skewed towards mosasaur body size, We then calculated the average PF size for each genera using Minitab version 17 (Fig. 4). We also determined the paleolatitudinal coordinates for each specimen from the Paleobiology Database Website (paleobiodb.org), with the highest known distribution for all genera presented in Table 3. 13

22 Figure 4. Average PF size among genera with standard deviation bars. A LSD test indicated there was significant differences between Clidastes, Mosasaurus, and Tylosaurus to Platecarpus (p<0.001) and there was a significant difference between Platecarpus to Plioplatecarpus (p<0.001) Mosasaur Genera Highest Southern Paleolatitude Distribution Highest Northern Paleolatitude Distribution Clidastes N/A 54.7 N (Sternberg, 1915) Mosasaurus 62.6 S (Fernández and Gasparini 2012) 57.1 N (Nicholls and Russell, 1990) Platecarpus 17 S (Bengtson and Lindgren, 2005) 53.7 N (Bardack, 1968) Plioplatecarpus 62.6 S (Fernández and Gasparini 2012) 78.2 N (pers comm. T. Konishi) Tylosaurus 64.6 S (Warren and Speden, 1978) 53.7 N (Bardack, 1968) 14

23 Table 3. Highest known northern and southern paleolatitudinal distribution for Clidastes, Mosasaurus, Platecarpus, Plioplatecarpus, and Tylosaurus. Note Clidastes is restricted to the Northern Hemisphere We performed both a nonphylogeny and a phylogeny analysis for H1. A phylogeny analysis was performed to determine if there was an evolutionary relationship between PF size among genera and paleolatitudinal distribution. For the nonphylogeny analysis, Minitab version 17 was used to perform a One-Way Analysis of Variance (ANOVA) and a follow up post hocs (Fisher s Least Significant Difference [LSD] test) to determine if the genera were significantly different (alpha=0.05) (Fig. 4). A linear regression model was then used to determine the relationship between PF size and paleolatitude for both among and within genera (H1; Fig. 5) (alpha=0.05). For the phylogeny analysis, R version (Fire Safety) (R Code Team) was used to perform a phylogenetic independent contrasts (PIC) (Felsenstein, 1985) analysis, while assuming a Brownian motion process, using a modified version of the mosasaur phylogenetic tree provided by Bell (1997) (alpha=0.05) (Fig. 6). R converted both the PF/PB ratio and paleolatitudinal distribution to their independent contrasts by way of R library ape (Paradis et al., 2004) and caper (Orme et al., 2013) using the crunch function and then calculated their correlation through the origin (Fig. 7). For H2, an ANOVA was used to test for PF size differences between taxa that lack avascular necrosis (i.e., Clidastes) versus genera that possess this character (i.e., Mosasaurus, Platecarpus, Plioplatecarpus, and Tylosaurus (Table 4). 15

24 Figure 5. Relationship between paleolatitudinal distribution and parietal foramen (PF) length over parietal bone length (PB) (PF/PB) among genera. All paleolatitudinal points were acquired from the Paleobiology Database Website (paleobiodb.org). A, graphical representation of the most extreme paleolatitudinal distribution for each mosasaur genera. The sources for each extreme paleolatitudinal point can be found on Table 3. B, Clidastes; C, Mosasaurus; D, Platecarpus; E, Plioplatecarpus; F, Tylosaurus are graphical representations of the relationship between paleolatitude and PF/PB for specimens within genera. Only specimens with known fossil locality information were used

25 Figure 6. A generalized mosasaur tree displaying the five main genera studied for the experiment along with their mean parietal foramen length/parietal bone length (PF/PB) ratio, highest paleolatitude known, and presence of avascular necrosis. Tree modified from Bell (1997) and avascular necrosis data came from Rothschild and Martin (2005). Branch lengths were set to be equal

26 Figure 7. Plot of the residual independent contrasts of parietal foramen length/parietal bone length against the residual independent contrasts of paleolatitude among the five mosasaur genera (R 2 =0.601, p=0.077). The line of regression intercepts at (0,0) Mosasaur Genera Clidastes Mosasaurus Platecarpus Plioplatecarpus Tylosaurus Presence of avascular necrosis no yes yes yes yes Table 4. Presence or absence of avascular necrosis in mosasaurs from Rothschild and Martin (2005) RESULTS Plioplatecarpus had the largest PF with an average PF/PB ratio of 0.43 (SD=0.06), followed by Platecarpus (PF/PB=0.18, SD=0.02), Mosasaurus (PF/PB=0.09, SD=0.02), Tylosaurus (PF/PB=0.07, SD=0.02), and Clidastes (PF/PB=0.07, SD=0.02) (Fig. 4). The ANOVA found statistically significant differences in PF size among genera (F=201.28, DFn=4, DFd=40, p<0.001). A follow up Fisher s LSD test determined three significantly different groups among the five genera: Group A (Plioplatecarpus), Group B (Platecarpus), and Group C (Clidastes, Mosasaurus, and Tylosaurus) (p<0.001). For H1, the nonphylogeny-oriented, linear regression model determined that there was a weak but nonsignificant correlation between PF size and paleolatitude when all genera were considered (R 2 =0.582, p=0.134) (Fig. 5A). The phylogeny-oriented PIC analysis falsified the hypothesis, as there was a weak but nonsignificant correlation 18

27 (R 2 =0.601, p=0.077) (Fig. 7) between contrasts in PF size and contrasts in paleolatitudinal distribution. Within genera, a linear regression model determined that Clidastes and Plioplatecarpus had a positive but negligible correlation between PF size and paleolatitude (R 2 <0.1), whereas Mosasaurus, Platecarpus, and Tylosaurus exhibited weak and negative correlations between PF size and paleolatitude (R 2 <0.25) (Fig. 5B F). An ANOVA falsified H2 as Mosasaurus and Tylosaurus, which displayed avascular necrosis symptoms, have a PF size similar to that of Clidastes (see above LSD test), which did not possess avascular necrosis (Table 4; Rothschild and Martin, 2005). Plioplatecarpus and Platecarpus also had significantly larger PF size than both Mosasaurus and Tylosaurus even though these four genera all displayed symptoms of avascular necrosis DISCUSSION Paleolatitudinal Distribution among Genera Both the nonphylogeny and PIC analyses falsified H1 as there is a weak but nonsignificant relationship between PF size and paleolatitudinal distribution. In the linear analysis, the correlation coefficient and p-value (R 2 =0.582, p=0.134) indicates a weak and nonsignificant correlation between PF size and paleolatitude. Furthermore, the PIC analysis indicates that the evolution of a large PF does not associate with the evolution of more extreme latitudinal distribution in mosasaurs (R 2 =0.601, p=0.077) (Fig. 7). Plioplatecarpus had by far the largest PF, as determined by a Fisher s LSD test, and the highest paleolatitudinal distribution (~78 N) (T. Konishi, personal communication, 2013) of the five genera studied. A large PF suggests that Plioplatecarpus had a large PE, which would likely have been more sensitive to low-intensity sunlight compared to such smaller PF- 19

28 bearing mosasaurs as Tylosaurus. Although Plioplatecarpus had the most northern distribution and the largest PF, it shared nearly equal southern paleolatitudinal extremes (~64 S) with such smaller PF-bearing mosasaurs as Mosasaurus and Tylosaurus (Fernández and Gasparini, 2012). Clidastes, Mosasaurus, Platecarpus, and Tylosaurus also had a similar northern paleolatitudinal distribution (~55 N) even though Platecarpus had a larger PF than Clidastes, Mosasaurus, and Tylosaurus (p<0.001 in Fisher s LSD test) (Sternberg, 1915; Bardack, 1968, Nicholls and Russell, 1990) (see Fig. 3 4, Table 3). If, in the future, additional specimens of Platecarpus are discovered at or described from latitudes > 55, correlation between PF size and paleolatitudinal distribution may be stronger and can push both the linear regression analysis (p=0.134) and the PIC analysis (p=0.077) p-values to significant values. There are already several partial remains of mosasaurs found in Australia that are tentatively described as Platecarpus (Kear et al., 2005). These specimens have a slightly higher paleolatitudinal range (~58 S) than the current specimens found in southern Canada (Bardack, 1986). Platecarpus specimens are also likely to be found at the Anderson River site in the Northwest Territories of Canada (see Russell, 1967) and could further extend their paleolatitudinal range to ~78 N. Given that Platecarpus are more abundantly found in the northern part of the Western Interior Seaway (Nicholls and Russell, 1990), finding a Platecarpus specimen at the Anderson River site would not be unlikely. If this were to happen, an additional PIC analysis could give us a significant p-value and reveal a correlation between the evolution of PF size and latitudinal distribution Paleolatitudinal Distribution within Genera 20

29 Hypothesis 1 within genera was falsified, as there is no apparent relationship between PF size and paleolatitudinal distribution for the specimens analyzed. There was a very weak to negligible correlation among the five genera ranging from R 2 =0.210 to (Fig 5B E). The PF size for each individual showed almost no association with its fossil locality. The weak correlation between PF size and paleolatitude indicates the absence of an association between PE and local latitudinal position. Alternatively, the lack of a relationship between PF size and paleolatitudinal distribution within mosasaur genera may, instead, reflect an adaptive value of the PE (i.e., PF by proxy) with a role related to the onset of migration behavior. Mosasaurs with a larger PE size may have a behavioral response to day length (sensed with the PE) and initiate migration across different latitudes, following the intensity of solar radiation. These movements may have taken such mosasaurs as Plioplatecarpus from northern Canada (e.g., Anderson River site in the Northwest Territories during the Late Cretaceous; ~78 N) to southern Alabama (e.g., Mobile area during the Late Cretaceous; ~30 N). This would allow the mosasaurs to maintain their circadian rhythm and body temperature as incident sunlight angle increased at lower latitude during shorter daylight hours experienced during the (present day) winter months. This explanation is based on homing and migrational behaviors exhibited by extant reptiles that are controlled by the pineal complex (e.g., Freake, 2001; Ellis-Quinn and Simon, 2004) and the study of the migrational effects of the pineal body on Dermochelys coriacea (Davenport et al., 2014). The thinning of the PB in D. coriacea is analogous to the PE in extant and extinct squamates, as both structures would allow a greater amount of sunlight to be absorbed by the pineal body. If the rather large and highly exposed pineal body in D. coriacea triggers migration via the relative 21

30 amount of day length, then, perhaps, mosasaurs with a large PE, like Plioplatecarpus, used their PE and pineal body for migration in the Western Interior Seaway during the Late Cretaceous Presence of Avascular Necrosis vs. PF size among Genera Hypothesis 2 is falsified, as there is no relationship between PF size and presence of avascular necrosis. If there was a relationship, we would have seen mosasaurs with a small PF (Clidastes, Mosasaurus, and Tylosaurus) not display characteristics of avascular necrosis while the opposite would hold true for mosasaurs with a large PF (Platecarpus and Plioplatecarpus). This was not the case as specimens of Clidastes did not exhibit avascular necrosis, whereas Mosasaurus, Platecarpus, Plioplatecarpus, and Tylosaurus did. Thus, there is no correlation between these two features in mosasaurs. Interestingly, Rothschild and Martin (2005) reported that Plioplatecarpus had a smaller percentage range of vertebrae affected by avascular necrosis (5 17%), whereas Platecarpus had a much greater percentage range of affected vertebrae (10 64%). Their study is consistent with our rejection of H2, as Plioplatecarpus has the significantly largest PF and would have a greater range of vertebrae affected by avascular necrosis when compared to Platecarpus Resolving the Function of the PF Although H1 explains Plioplatecarpus and its significantly large PF, the purpose of the large PE in Platecarpus versus other mosasaurs (p<0.001 in Fisher s LSD test) is still unclear (see Figs. 2, 4). Neither H1 nor H2 explains the size difference versus paleolatitude or presence of avascular necrosis. 22

31 To resolve the function of the PE in Platecarpus and other mosasaurs, we encourage further fieldwork to discover mosasaur-bearing strata as they may lead to new discoveries of previously unidentified mosasaurs, fill in the paleobiogeographical gaps in the fossil record (e.g., Northwest Territories in Canada, Australia, Antarctica), and produce more specimens with intact PB. Although there is a particular interest in the paleobiogeographic range of the five genera analyzed here, other mosasaur genera measured for this study are also of interest, such as Prognathodon and Tethysaurus. Additional genera will be beneficial in performing a follow-up PIC analysis, as this may provide a stronger, more robust understanding of the evolutionary relationship between PF size and latitudinal distribution of mosasaurs. The greatest challenge in determining the true paleobiogeographical range of mosasaurs is the extreme high-latitude field sites, as they are located in either inhospitable or highly isolated locations. Mosasaurs from these sites would provide insight into whether the PF was different in size or shape for high-latitude dwelling mosasaurs compared to their low-latitude dwelling brethren. The Northwest Territories of Canada are of particular interest, as only a mosasaur quadrate has been recovered from one field site, located near the Anderson River, as a result of a single excavation (Russell, 1967). The quadrate bone (NMC 10429) was originally identified as Platecarpus but is now identified as Plioplatecarpus (T. Konishi, personal communication, 2013). This reidentification decreased the known distribution for Platecarpus to southern Manitoba (paleolatitude of ~53 N) placing it with such smaller PF-bearing mosasaur genera as Tylosaurus and Clidastes. Antarctica has also yielded the mosasaur genus Taniwhasaurus a close relative of Tylosaurus from James Ross Island (~62 S) and Vega Island (~60 S) (Fernández and Gasparini, 2012). Although Tylosaurus, itself, has yet to be 23

32 identified in Antarctica, two vertebrae (MLP 87-II-7-1) have been found on the nearby Seymour Island (~62 S) and are identified as tylosaurine (Fernández and Gasparini, 2012). Other high-latitude mosasaur sites are more accessible but have mostly yielded fragments and partial remains of teeth and vertebrae. Western Australia, near Gingin, Dandaragan, and the Giralia Ranges, is a prime location as supposed remains of Platecarpus have been found in the Molecap Greensand (~58 S). These remains, however, lack a skull, are in poor condition, and have yet to be confidently identified (Kear et al., 2005). New Zealand sites have a similar paleolatitude (~58 S) and contain tylosaurine (Caldwell et al., 2005), mosasauridae (Wiffen, 1990), and other mosasaur specimens (Wiffen, 1990 and Consoli and Stilwell, 2009). Although mainly partial fragments have been identified a single quadrate bone representing the Mosasaurus specimen (Wiffen, 1990) two Taniwhasaurus oweni specimens (NMNZ R 1536 and KHM N ) have been identified by their jugal, frontal, prefrontal, premaxilla, pterygoid, and even a PB (Caldwell et al., 2005). If more specimens of Taniwhasaurus with their PB intact are discovered in New Zealand and Antarctica, a new PF-analysis could be undertaken on this genus. This analysis would be particularly interesting as Taniwhasaurus has been constantly found in high-latitude field sites. The PF size of Taniwhasaurus could be compared to the PF size of its lower latitude cousin, Tylosaurus, via H1 in another PIC analysis. Japan (~42 N) has yielded mosasaurs relevant to this study, namely Mosasaurus (see Sato et al., 2012 for general occurrence of mosasaurs in Japan). The Mosasaurus specimens are represented mainly by teeth, vertebrae, and other skeletal fragments, but we are hopeful that more remains, especially their PB, will be extracted from the field sites. The Kristianstad Basin in Southern Sweden (~46 N) is another site in which partial 24

33 mosasaur remains of Clidastes (Lindgre and Siverson, 2004) and Tylosaurus (Lindgren and Siverson, 2002), have been found. If enough viable specimens from these two genera were recovered, we could observe if there is any difference in PF size between European and North American specimens within the same genera. Mosasaurs found in lower latitude sites could be used to retest H1 for within genera. Mosasaurs have been identified, such as Yaguarasaurus and Tethysaurus, in Columbian (~3 S) deposits (see Páramo-Fonseca, 2011 for a general review of their occurrences) and a Columbian Campanian mosasaur (ING RC090805) has yet to be confidently identified as it is currently under preparation (Bengtson and Lindgren, 2005). Two isolated Platecarpus tooth-crowns have also been found in the Sergipe Basin in northeastern Brazil (~17 S) (Bengtson and Lindgren, 2005). The Oulad Abdoun Basin in Morocco (~25 N) has also produced an abundance of mosasaurs, as well as a recent number of publications identifying and describing new mosasaurs (e.g., Bardet et al., 2004, 2005a, 2005b; Leblanc et al., 2012) including specimens of Mosasaurus. Although no Mosasaurus specimens have yet been found with intact PB, two Eremiasaurus specimens (UALVP and OCP DEK/GE 112) were both preserved with a nearly complete skull and vertebral column (Leblanc et al., 2012). Comprehensive studies of PE size (PF size and its skull position as proxies) in extant lizard taxa with respect to their home range, latitude, altitude, and climate variables would be useful to understand PF size variation in PE-bearing ancient animals. Only a few studies (e.g., Gundy et al., 1975 and Labra et al., 2010) have intensely analyzed these relationships so the amount of new information that could be obtained would greatly benefit paleontological research. This could be accomplished by measuring PE, PF, and pineal 25

34 body size amongst different groups of closely related reptiles who live in vastly different environments or latitudes such as Liolaemus (Labra et al., 2010). There could also be other factors that affect PE size in modern reptiles that we are not aware of yet that could explain why Platecarpus has a larger PE than Clidastes or Tylosaurus. Lastly, we encourage further research on the relationship between the pineal body and seasonal migration in migratory sea turtles for both low-latitude seas turtles e.g., the loggerhead sea turtle (Caretta caretta) and the green sea turtle (Chelonia mydas) and such high-latitude sea turtles as the leatherback sea turtle (Dermochelys coriacea). The methodology in the study of the leatherback sea turtle by Davenport et al. (2014) should be followed to deduce the relationship between latitudinal distribution, water temperature, and turtle movement throughout the year. If there is a relationship between migration and pineal activity, this information could be applied to mosasaurs to test the new hypothesis proposed here that the PE was used to trigger and control the latitudinal range of migration CONCLUSIONS We tested two hypotheses for this study concerning the PF of mosasaurs (Table 1). We found that there is a weak but nonsignificant relationship between PF size and paleolatitudinal distribution among mosasaur genera. Plioplatecarpus has both the largest PF and the highest known distribution of mosasaurs. The mosasaur with the second largest PF, Platecarpus, however, shared a similar paleolatitudinal extent with other, much smaller PF-bearing mosasaurs, such as Clidastes, Mosasaurus, and Tylosaurus. There was also a nonsignificant relationship between PF size and paleolatitude distribution within mosasaur genera. We also determined that there was no relationship between the presence of 26

35 avascular necrosis and PF size. As such, we reject both hypotheses. We propose a new hypothesis that the PE (PF as the proxy) was used for migrational behaviors, as the PE could be a used as a trigger for seasonal migration due to changes in day length and incident angle of sunlight. The substantial difference in PF sizes among Platecarpus and other mosasaur genera (p<0.001 in Fisher s LSD test) warrants future investigation into why Platecarpus evolved a large PE. This study can be used as a framework to test the function of the PE, with the PF as a proxy, in other extinct vertebrates, such as the marine plesiosaurs (Lepidosauromorpha) and the terrestrial therapsids (Synapsida) ACKNOWLEDGEMENTS We thank Dr. Konishi for reidentifying a key specimen that was found in Northwest Territories 50 years ago and was almost completely forgotten. Thanks to Dr. Mike Everhart and Dr. Michael Polcyn for discussions that improved the quality of this manuscript. We thank Gil Parker, president of the Kansas City Paleontology Club, for coordinating with the Black Hills Institute in South Dakota on the behalf of the authors. We also thank the Alabama Museum of Natural History, the Fort Hays Sternberg Museum, the Field Museum of Natural History, the University of Kansas Museum of Natural History, the Muséum National D'histoire Naturelle, the Institut Royal des Sciences Naturelles de Belgique, the Mcwane Center, the Shuler Museum of Paleontology, the Royal Tyrell Museum of Paleontology, National Museum of Natural History and their respective curators for access to their collections. Thanks to the University of Kansas IchnoBioGeoScience (IBGS) research group for discussion of earlier drafts of this manuscript. 27

36 REFERENCES Bardack, D Fossil vertebrates from the marine Cretaceous of Manitoba. Canadian Journal of Earth Sciences 5: Bardet, N., X. Pereda Suberbiola, M. Iarochene, B. Bouya, and M. Amaghzaz. 2005a. A new species of Halisaurus from the Late Cretaceous phosphates of Morocco, and the phylogenetical relationships of the Halisaurinae (Squamata: Mosasauridae). Zoological Journal of the Linnean Society 143: Bardet, N., X. P. Suberbiola, M. Iarochène, M. Amalik, and B. Bouya. 2005b. Durophagous Mosasauridae (Squamata) from the Upper Cretaceous phosphates of Morocco, with description of a new species of Globidens. Netherlands Journal of Geosciences 84: Bardet, N., X. P. Suberbiola, M. Iarochene, F. Bouyahyaoui, B. Bouya, and M. Amaghzaz Mosasaurus beaugei Arambourg, 1952 (Squamata, Mosasauridae) from the late Cretaceous phosphates of Morocco. Geobios 37: Bell, G A phylogenetic revision of North American and Adriatic mosasauroidea; pp in J. Callaway and E. Nicholls (ed.), Ancient Marine Reptiles. Academic Press, Cambridge, Massachusetts. Beltrami, G., C. Bertolucci, A. Parretta, F. Pertrucci, and A. Foa A sky polarization compass in lizards: the central role of the parietal eye. Journal of Experimental Biology 213: Bengtson, P., and J. Lindgren First record of the mosasaur Platecarpus Cope, 1869 from South America and its systematic implications. Revista Brasileira de Paleontologia 8: